Alcohol dependence (AD) is a common problem with significant health consequences. Treatment of AD is evolving to include both counseling methods and medications. Several medications have been discovered, that show efficacy in AD, e.g. naltrexone, acamprosate. However, the overall effect of existing medications is modest leaving a clear need for the development of new pharmacotherapies. The GABAB receptor agonist baclofen has attracted attention as a potential new medication for AD based on preclinical data and early clinical trials. Baclofen is an FDA approved medication with an excellent safety profile even for patients with liver cirrhosis-a not uncommon consequence of AD. Questions have arisen with regards to the efficacy of baclofen and whether higher doses of baclofen are safe and more effective than the prior tested dose of 30 mg/ day. There is emerging evidence that severity of dependence is positively associated with baclofen response. The main goal of the present proposal is to test the efficacy and safety of 30 mg/d and 90 mg/d of baclofen compared to placebo controlling for severity of dependence as assessed by drinks/drinking day. A primary secondary goal will examine for an anxiolytic effect of baclofen. The study proposes to enroll 120 men and women with AD in a randomized, placebo-controlled trial to include at least 60 individuals with more severe AD (e14 drinks/drinking day for men; e10 drinks/drinking day for women) with randomization to baclofen or placebo balanced for this variable. Baclofen will be titrated to 10 mg t.i.d over 3 days and to 30 mg t.i.d over 12 days and maintained at that level for 12 weeks and then down- titrated for a total study time of 16 weeks. Medical Management will be provided to encourage progress towards drinking goals and to enhance retention and compliance. Drinking patterns, anxiety levels, sleep patterns, craving for alcohol, gamma-glutamyl transferase (GGT) and carbohydrate deficient transferring (CDT) will be assessed. Trough blood levels of R & S-baclofen will be assessed in all individuals at week 4. In summary, the present proposal is innovative and of clinical significance as it will test and compare standard and high-dose baclofen for efficacy and safety in individuals with AD. The proposal is adequately powered to test the primary hypothesis and provides good power to assess whether drinks/drinking day is predictive of baclofen response. Adequate power is also present to examine the anxiolytic effect of baclofen. Ascertaining the effects of standard and high-dose baclofen, the predictive value of heavy drinking on baclofen response and the anxiolytic effect of baclofen are important goals towards determining whether baclofen has true value for the clinical management of the patient with alcohol dependence. .